Closed natolambert closed 7 months ago
Solution for the relevant models (will need more filtering code)
############################
# Tokenization settings & dataset preparation
############################
# set pad token to eos token if not set
if reward_pipe.tokenizer.pad_token_id is None:
reward_pipe.model.config.pad_token_id = reward_pipe.tokenizer.unk_token_id
reward_pipe.tokenizer.pad_token_id = reward_pipe.tokenizer.unk_token_id
# make tokenizer add eos
reward_pipe.tokenizer.add_eos_token = True
TLDR:
Seems like FastChat models are not getting an EOS token. Could effect the non-DPO models that are using FastChat chat templates minorly